Ficaria verna , and other Members of the Ranunculaceae . 7 1 
but an expression of the various mechanical forces in the developing bud, 
and can give us no clue to the primitive construction of the flower ; but the 
order of dehiscence and the arbitrary recurrence of certain numbers cannot 
be explained on purely mechanical grounds and must be construed as of 
phylogenetic import. 
(b) Aconitum Lycoctonum. The androecium of this species was exam- 
ined in forty unopened flower-buds obtained from wild plants in Switzerland 
by Prof. F. E. PTitsch. From the variation ‘curve’ given in Fig. 18 it will 
be seen that the range is narrow and that in the majority of cases the number 
of stamens was thirty. The lower limit observed was twenty-nine, but 
Eichler ( 1875 , p. 164, Fig. 67) figures a specimen with only twenty- seven 
stamens. The large proportion of examples with thirty-three stamens is 
strongly suggestive of a second maximum. In all the flowers examined the 
gynaeceum consists of three carpels, so that the trimerous tendency is here 
even more marked than in its congener, a fact that may not be without 
significance in view of the greater specialization of A. napellus. 
V. Aquilegia vulgaris, L. 
We have assumed that where a frequency • curve ’ shows maxima at 
multiples of three the explanation lies in an innate trimerous tendency, as 
a consequence of which the variation ‘ curve’ is to be regarded as composed 
of a series of subsidiary variation ‘curves’ around the different multiples of 
three, each, however, of diminished prominence in proportion as it departs 
from the primary mode. 
It is well known that the flower of Aquilegia is whorled throughout and 
is typically pentamerous in structure, and it was thought that the variation 
in the number of the carpels might afford corroboration of the above assump- 
tion by the presence of maxima at multiples of five. For this purpose 
300 flowers were examined, and it was found that two maxima were present, 
namely, at five and ten (cf. Table IX). Even here, however, the lower limit 
was three carpels. 
Table IX. 
Aquilegia vulgaris . 
Number of Carpels. Number of Specimens. 
3 
3 
4 
3 
5 
202 
6 
30 
7 
22 
8 
13 
9 
6 
10 
20 
11 
0 
12 
0 
13 
1 
As in the essentially trimerous flowers we have been considering the 
stamens are often not a multiple of three, so too here the number of stamens 
